Multi Reviewer Patterns
Use when coordinating independent code reviews across security, performance, architecture, testing, accessibility, or reliability dimensions and consolidating findings without losing evidence or dissent.
Install
npx skills add sandbaseai/workbuddy-skill --skill multi-reviewer-patternsgit clone --depth 1 https://github.com/sandbaseai/workbuddy-skill.git /tmp/workbuddy-skill
mkdir -p ~/.claude/skills
cp -r /tmp/workbuddy-skill/skills/multi-reviewer-patterns ~/.claude/skills/multi-reviewer-patternsPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# 多评审者协同模式 多评审者评审的价值在于独立视角和可比较证据,不在于把更多意见相加。先按变更风险分配维度,再让每个评审者从同一 base/head 范围独立工作,最后去重、校准严重度、保留分歧并形成可行动报告。 ## 范围与安全门禁 开始前记录仓库、精确 base/head、文件范围、授权读者、是否允许访问依赖/测试环境、报告保留期和合并权限: - 评审只读取授权代码、配置、测试和证据;不把私有代码、秘密、客户数据或内部 URL复制到评审上下文或报告; - 评审者之间默认隔离,不共享未整理的结论,避免早期意见造成锚定; - 发现必须绑定精确文件/行号或可复核范围、触发条件、影响和证据;没有证据的内容标记 `Hypothesis` 或不纳入结论; - 多数票、作者身份、代码风格偏好或同一行出现多次不能单独证明严重度; - 评审结果不自动修改代码、关闭线程、批准合并或发布;这些动作需要独立授权和门禁。 目标范围变化、base/head 不稳定、关键文件缺失或敏感材料无法脱敏时返回 `BLOCKED`。 ## 维度分配 | 维度 | 关注点 | 典型触发 | | --- | --- | --- | | Security | 信任边界、认证、授权、输入、秘密、供应链 | 用户输入、权限、外部集成 | | Correctness/Architecture | 不变量、边界、耦合、接口、迁移 | 新模块、跨层改动、架构变化 | | Performance/Reliability | 查询、资源、并发、超时、降级 | 热路径、依赖、容量或并发变化 | | Testing | 行为覆盖、负面路径、隔离、回归 | 新功能、修复、风险较高的重构 | | Accessibility/UX | 键盘、语义、ARIA、反馈和错误恢复 | UI、表单、动态内容 | 只选择与变更风险相关的维度,并说明未选择维度的理由。一个评审者可以承担多个低冲突维度,但高风险安全结论最好由独立上下文复核。 ## 独立评审流程 1. **固定范围**:记录精确 base/head、生成 diff、文件列表、构建/测试状态和已知限制。 2. **创建任务卡**:为每个维度写关注问题、排除范围、截止时间、证据格式和停止条件。 3. **独立阅读**:先读需求/不变量和调用边界,再读变更;不读取其他评审者的 finding 草稿。 4. **验证发现**:检查调用者、错误路径、测试、配置和依赖上下文;能运行时只在授权隔离环境执行最小验证。 5. **标准化 finding**:每项包含标题、位置、维度、事实、触发条件、影响、证据、严重度、建议和不确定性。 6. **去重与关联**:按“同一问题”而非仅按同一行合并;保留不同维度的独立风险和所有支持证据。 7. **校准严重度**:结合影响、可利用性/发生概率、暴露范围、可检测性、可逆性和现有控制;记录为何高于或低于初始评级。 8. **处理冲突**:对矛盾事实、严重度或建议保留双方依据,指定复核动作,不用投票静默覆盖。 9. **汇总与门禁**:按严重度和证据排序,给出整体结论、未决项、合并前阻断项和安全下一步;报告完成不代表代码已合并。 ## Finding 模板
Details
- Skill name
multi-reviewer-patterns- Repository
- sandbaseai/workbuddy-skill
- Path
skills/multi-reviewer-patterns/SKILL.md- License
- MIT
- Repo updated
- 8 September 2026
FAQ
What does the Multi Reviewer Patterns skill do?
Use when coordinating independent code reviews across security, performance, architecture, testing, accessibility, or reliability dimensions and consolidating findings without losing evidence or dissent.
How do I install the Multi Reviewer Patterns skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add sandbaseai/workbuddy-skill --skill multi-reviewer-patterns. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Multi Reviewer Patterns popular?
It lives in sandbaseai/workbuddy-skill, which has 9 GitHub stars and 3 forks.
Related skills
Access Control Security Best Practices
9sandbaseai/workbuddy-skill
Design, review, implement, and verify least-privilege access controls across users, services, tenants, resources, and administrative workflows.
Accessibility Review
9sandbaseai/workbuddy-skill
Use when reviewing a web interface or interaction for semantic, keyboard, focus, screen-reader, ARIA, and contrast accessibility risks.
Acquire Codebase Knowledge
9sandbaseai/workbuddy-skill
Use when the user asks to map, document, or onboard into an existing codebase and needs evidence-backed architecture, structure, conventions, integrations, testing, and concerns documentation.
Multi Source Knowledge Library
9sandbaseai/workbuddy-skill
Use when several books, documents, web pages, or repository files should become one searchable, cross-referenced knowledge library instead of disconnected summaries.
Mobile Development
9sandbaseai/workbuddy-skill
Use when building or reviewing iOS, Android, React Native, Flutter, SwiftUI, or Jetpack Compose applications; choose architecture and platform boundaries while designing for unreliable networks, battery, memory, accessibility, security, testing, and staged release.
Oauth Debugging
9sandbaseai/workbuddy-skill
Use when an OAuth 2.0 or OpenID Connect flow fails, loops, loses session state, rejects a token, or grants the wrong access; reconstruct the exact protocol exchange without exposing credentials or live tokens.
Browse the most popular Claude skills, all software development skills, or write your own with the skills generator.